SmartTrader V4

Inspiration

What it does

Our application is a real-time AI crypto currency trader driven by sentiment and technical analysis. Each day we track both market sentiment, technical indicators and use the overall trend to predict buy or sell indicators.

How we built it

We use two different clever approaches to collect data. We have access to a large dataset from X, which contains information about Bitcoin public sentiment. We use this large dataset to populate 6 months of public sentiment for our SmartTrader to backtest. To collect real-time data, we take advantage of Reddit’s API, to take advantage of detailed, deep conversations to take swing trading positions (over days to weeks) . By taking advantage of VADER, the lightweight lexicon based nature allows us to efficiency and accurately categorize data between -1 and 1, with -1 representing a negative sentiment, and 1 representing a positive sentiment. By implementing Alpaca’s powerful API, we can get access to real-time market data for Bitcoin to properly assess moving averages and Relative Strength Index. We cross check these two model’s to see if they point to a buying signal, then cross check with our sentiment analysis model to then come out to a final verdict. Front-End: React, Daisy UI, Tailwind, Chart.js Back-End: Python, Fast API, PRAW, VADER, Alpaca Trade API, PostgreSQL, Pandas By combining a lot of these lightweight powerful tools, our application is able to efficiently judge, and enter positions.

Challenges we ran into

We had difficulties with financial side of application, specifically understanding financial approach in regards to number of queries we had access to, targeted slightly volatile markets as a solution.

Accomplishments that we're proud of

We are proud of our product being finished, and the fact that we battled through a completely new topic.

What we learned

We learned more about how quantitative firms operate, algorithmic programming, low-latency programming, machine learning, and full-stack development with a python back-end.

What's next for SmartTrader V4

Though our product only supports Bitcoin, we plan on applying this to most CEX cryptocurrencies, and plan on targeting smaller caps to test our product in more volatile markets.

Built With

• fastapi
• postgresql
• python
• react
• tailwindcss
• typescript
• vader